Aid effectiveness 1

By Finn Tarp

Director, UNU-WIDER

1. Introduction

The question of foreign aid’s impact on economic growth is highly controversial and excites polarised opinions. Optimistic views of aid are exemplified by those of Jeffrey Sachs (2005), who calls for a doubling of worldwide aid flows as a moral obligation of rich countries that will send ‘forth mighty currents of hope’ and lead to ‘the end of poverty’. In contrast, William Easterly (2001, 2003, 2006) is a vocal sceptic, highlighting aid’s apparent historic inability to buy growth. And today, in the midst of a serious global economic crisis, where aid is arguably more needed than ever, the attention of both the aid community and decision- makers is on “Dead Aid” (Moyo, 2009), which argues for a complete cessation of aid flows to Africa.

This note provides an overview of the academic literature that has evaluated the effectiveness of aid across countries. Focus is on the aggregate impact of aid; that is, its effect on country- wide indicators, particularly economic growth. Sections 2 and 3 provide some essential background to these debates, while Section 4 provides an answer to the core question – does aid support economic growth? In doing so, the note draws on a review of both microeconomic and macroeconomic evidence, spanning four generations and more than 40 years of literature.

The question in focus is not as simple as it may appear. Methodological difficulties mean that it is hard to provide a definitive and comprehensive conclusion. Even so, there is very little to suggest that foreign aid has a consistent harmful effect across countries over time, and a balanced view of the literature and the data indicate that foreign aid does provide growth and development benefits to recipients. These benefits are not as large as sometimes argued and they accrue over considerable time frames, 30 years or more. Thus, aid should never be

1 This note was prepared based both on joint work with Channing Arndt and Sam Jones including an unpublished report for NORAD and our 2009 UNU-WIDER research paper entitled “Aid and Growth: Have We Come Full Circle?” dated September 2009. It can be downloaded from http://www.wider.unu.edu/publications/working-papers/discussion-papers/2009/en_GB/dp2009-05/ . I am grateful for helpful comments from many colleagues including Erik Thorbecke.

1 considered to be a ‘solution’ in itself to poverty. Section 5 summarises these lessons from the literature; and Section 6 concludes.

2. Why are some countries poor?

The disparities in living standards across countries in the world today are spectacularly large. Using per capita income figures, the average citizen of a wealthy country consumes about as much in a few days as the average citizen of many poor countries consumes in a year. Adjusted for differences in purchasing power, the gap narrows somewhat. Even so, the differences are so large that they defy precise measurement. Rough adjustments for purchasing power imply differentials in per capita income/consumption between wealthy and poor countries of about 50 times. Large differences also exist using other measures of well- being such as life expectancy, educational attainment, and infant mortality.

The proximate cause of this chasm in living standards relates to productive capacity. Poor countries simply do not produce very much. Economists typically view the productive capacity of an economy as being a function of stocks of factors of production especially physical capital (factories, tools, computer, roads etc.), human capital (workers with skills) and labour (workers without skills) as well as technology. Economists also posit a crucial role for institutions, which provide frameworks for organizing these factors in a manner that is productive. Poor countries lack physical capital, human capital, technology, and well functioning institutions. As a result of the lack of these four attributes, the mass of unskilled labour present in the country is highly unproductive with negative implications for nearly all aspects of well-being.

There are some common elements to these four attributes that help to explain their relative absence in poor countries. First, they all accumulate through long run processes. The buildings in most developed country cities reflect more than a century of cumulative construction. Converting a kindergartner into a highly skilled engineer takes at least two decades. While technological leaps are possible, technology improvement is mainly a slow evolutionary process conducted via repeated trial and error. Finally, institutions evolve slowly over time. For example, British common law is based on centuries of case experience.

Second, accumulation of these attributes generally requires a forward looking mind set. In order to accumulate any of these, one must typically sacrifice something today in order to benefit in the future. This can be very difficult indeed when current resources are highly

2 constrained and when people are unable to satisfy their basic needs. In this case there is no scope for saving. Third, accumulation of these attributes relies on public private partnerships. Even strongly market oriented economies, such as the United States, rely on the public sector to supply basic economic infrastructure such as roads and bridges, education, and to fund research and development for new technologies. This highlights the important role of institutions both for current and future production levels.

Finally, note that, unless very well established, all four of these accumulation processes are highly vulnerable to disruption. At the extreme, years of effort can be wiped out in short order. As a prominent example, wars destroy both physical and human capital. More subtle disruptions than the eruption of violence can also have serious impacts. Experience indicates that nascent institutions are particularly vulnerable to disruption. Even a relatively short period of neglect can substantially harm long run efforts at institution building.

In summary, to improve living standards significantly, poor countries must produce more – much more. To produce more, poor countries must initiate and maintain long run cumulative processes to build physical capital and human capital, acquire technology, and nurture institutions that facilitate growth. The role of aid for development, broadly conceived, is to support these long run cumulative processes. The success of the aid enterprise in accomplishing this objective is the focus of this note. However, before reviewing the debate about whether aid ‘works’ from a broad macroeconomic perspective, it is helpful to start with some background about aid. The next section provides a definition of foreign aid, describes how it has changed over time, and indicates recent trends.

3. What is foreign aid?

3.1. Definitions

Foreign aid takes many forms. The most important in terms of its size and influence is official aid. The definition of official development assistance (ODA) is provided by the Development Assistance Committee (DAC) of the Organization for Economic Cooperation and Development (OECD). ODA is calculated as the sum of grants and loans to aid recipients that: (a) are undertaken by the official sector of the donor country; (b) have as the main objective the promotion of economic development and welfare in recipient countries; and (c) are on concessional financial terms (i.e., with a grant element equal to at least 25 per cent of the total). In addition to these financial flows, technical co-operation costs are included in

3

ODA; but grants, loans and credits for military purposes are excluded. Transfer payments to private individuals, donations from the public, commercial loans and foreign direct investment (FDI) are not counted. Moreover, while it is common to treat ODA and foreign aid as the same thing, this is misleading. Assistance funded by non-governmental organizations (NGOs), which is foreign aid but not ODA, has grown very significantly in the last 25 years and now equals about one-third of official assistance.

The above definition indicates that foreign aid is not always a free resource transfer and often arrives with economic and political conditions. In many cases, official donors require that recipient countries pursue reforms or policies that the donors feel should promote economic growth or development, and aid may also be tied which reduces its effectiveness. With respect to the multilateral development finance institutions, such as the IMF and World Bank, these requirements are often known as stabilization and structural adjustment conditions. However, as discussed below, these conditions have been the source of much controversy. Some scholars have argued that the very policy reforms suggested by these institutions have been counterproductive and caused economic and social decline rather than growth.

One should also recognise that aid is often given for very different purposes – e.g., building infrastructure, expanding access to education, or responding to humanitarian emergencies. It is also the case that aid may be given primarily with an eye to the interests of donors (or firms from donor countries) rather than the recipients. In a review of aid allocation patterns, for example, McGillivray (2003) finds that past colonial links and political alliances are major determinants of foreign aid and that such strategic factors are at least as important as variables which reflect recipient needs (also see Alsina and Dollar, 2000). At the same time, and as the next section demonstrates, official aid flows have not been static. Very different motives have driven aid relationships over time, particularly as global conditions and dominant ideas have changed.

3.2. Foreign aid in historical perspective

In its modern form, foreign aid has its origins in the early 1940s and intensified after the disruption that followed the Second World War. The international economic system had collapsed, and war-ravaged Europe faced a critical shortage of capital and an acute need for physical reconstruction. The response was the European Recovery Program, commonly known as the Marshall Plan. During the peak years the USA transferred some 2-3 per cent of its national income to help restore Europe (approximately 20 – 30 times the level of foreign

4 assistance that the US provides today as a share of national income). The Marshall Plan, which was administered by the Organization for European Economic Co-operation (OEEC), the predecessor of the OECD, was highly successful.

It is useful to highlight distinctions between post-war Europe and the developing countries that have been the focus of foreign aid for the past four decades. In the aftermath of the Second World War, Europe primarily lacked physical capital. To be sure, the War killed a great many highly skilled people, damaged some institutions, and forced institutional change. Nevertheless, compared with the developing countries, post-war Europe was very well endowed with skilled people and institutional frameworks. Furthermore, before and during the War, European countries operated at the leading edge of technology. Hence, the fundamental missing element was capital, which foreign aid provided. The impressive results achieved by the Marshall Plan fuelled highly optimistic expectations about the future effectiveness of foreign aid.

After the success of the Marshall Plan, the attention of industrialized nations turned to the developing countries, many of which became independent during the 1960s. Economic growth in a state-led planning tradition became a key objective during the 1950s and 1960s, and it was widely believed that poverty and inequality would be quickly eliminated through growth and modernization (‘trickle-down’). While the Marshall Plan was built around support to finance general categories of imports and strengthen the balance of payments (i.e., program aid), from the early 1950s project aid became the dominant aid modality. Some donors continued to supply program aid, but aid was increasingly disbursed for the implementation of specific capital investment projects and associated technical assistance.

The multilateralism of aid became more pronounced after the mid 1970s. Multilateral channels were at the time seen as more efficient and less political than bilateral aid, so the UN, World Bank, and other multilateral agencies expanded their activities quite considerably. The 1970s also saw an increased focus on employment, income distribution, and poverty alleviation as essential objectives of development and foreign aid. The effectiveness of trickle-down was widely questioned, and new strategies referred to as ‘basic human needs’ (BHN) and ‘redistribution with growth’ (RWG) were formulated and propagated alongside more radical ‘dependency’ theories of development. Nevertheless, the typical project aid modality remained largely unchanged.

5

The ‘golden era’ of the 1960s and 1970s came to an abrupt end at the beginning of the 1980s. The (second) oil shock in 1979 reversed economic conditions, and there was a large increase in interest rates due to economic stabilization policies adopted by developed countries. The international debt crisis erupted, and macroeconomic imbalances were widespread among developing countries. On the political scene Ronald Reagan and Margaret Thatcher came to power in the US and the UK respectively, and at the World Bank Anne Krueger became Vice President and Chief Economist, replacing Hollis Chenery. Economic circumstances in the developing countries and the relations between the North and South changed radically. The crisis hit hard, especially in many African countries. Focus in development strategy and policy shifted to internal domestic policy failure, and achieving macroeconomic balance (externally and internally) became widely perceived as an essential prerequisite for renewed development. In addition, trade not aid became the dominant slogan among many leaders and economists.

During the 1980s, the goal of reducing the role of the state in the economy turned into a rallying call. Reliance on market forces, an outward orientation, and the role of the private sector, including NGOs, were emphasized by the World Bank and others. In parallel, poverty alleviation slipped out of view in mainstream agendas for economic reform, but remained at the centre of attention in more unorthodox thinking such as the ‘adjustment with a human face’ approach of UNICEF. At the same time, bilateral donors and international agencies struggled with how to channel resources to the developing world. Quick-disbursing macroeconomic program assistance, such as balance of payments support and sector budget support (which were not tied to investment projects, and which could be justified under the headings of stabilization and adjustment), appeared an ideal solution to the dilemma of maintaining the resource flow and the desire to promote policy reform. Financial program aid and adjustment loans (and eventually debt relief) became fashionable and policy conditionality more widespread. In other words, a rationale, which corresponded well with the orthodox guidelines for good policy (known as the ‘Washington Consensus’), had been found for maintaining the aid flow.

By the early 1990s, a number of factors had fomented rising scepticism of the benefits of foreign aid. These included: (a) the persistence of economic crises in much of the developing world; (b) geo-political changes following the end of the Cold War, which weakened patron- client relationships between developing countries and former colonial powers; (c) the perception that policy conditionality was failing to promote genuine policy reform; and (d) a

6 growing fear that aid was generating undesirable dependency relationships. Consequently, the traditional support given to foreign aid by vocal interest groups in the industrial countries receded. Bilateral and multilateral aid institutions were subjected to criticism, and at times characterized as blunt instruments of commercial interests in the industrial world or as self- interested, inefficient, rent-seeking bureaucracies. Moreover, acute awareness in donor countries of cases of poor governance, corruption, and ‘crony capitalism’ led to scepticism about the credibility of aid recipients.

Nevertheless, a resurgence of official development assistance, at least at the level of political rhetoric, occurred. During the late 1990s a consensus emerged among developed country governments that a renewed approach to development assistance, which avoided the mistakes of the past, was warranted. In part this stemmed from a recognition of the growing magnitude of global interdependencies (globalization) which meant that poverty and conflict in one corner of the globe could have implications for all. These views gave birth to the Millennium Development Goals (MDGs), directed explicitly at poverty reduction and addressing basic human needs such as access to education, health and water. The main focus was on the poorest countries and Africa in particular. In this vein, the (then) British Prime Minister Tony Blair launched his Commission for Africa in February 2004, publishing its findings in 2005. Other initiatives, such as the Jubilee Debt Campaign to cancel third world debts, received a high profile. Thus, the plight of the poor has become a part of popular discourse and a cause of celebrities and politicians alike.

3.3. Trends in foreign aid

In accordance with changes in motives and perceptions, aid volumes have fluctuated over time. Total aid grew steadily in real terms until the early 1990s. After 1992, however, total aid flows started to decline in absolute terms until the turn of the millennium. Since then, there has been a modest upturn in aid flows, but this is not on the scale that corresponds either to estimates of the funds deemed necessary to meet the MDGs or to the promises made by developed country governments (such as those of the G8 at the Gleneagles summit in July 2005). 2

Two further points merit comment. First, while aid flows have only grown slowly over the past decade, and in many cases have fallen in per capita terms due to consistent population

2 For further information on the commitments made at the Gleneagles summit see, for example: www.g8.utoronto.ca/summit/2005gleneagles

7 growth, private flows expanded considerably until the present financial and economic crisis. Thus, despite the widespread perception that foreign aid amounts to a very significant resource, it is now the case that in many developing countries official assistance is dwarfed by other external capital inflows such as foreign direct investment (FDI) and remittances. However, this phenomenon does not apply equally to all countries. For many of the poorest countries, particularly those in sub-Saharan Africa, foreign aid remains highly significant and is a vital source of funds to support import purchases and government spending.

Secondly, and as a consequence of the foregoing, the allocation of aid (as measured by the aid to GNI ratio) is highly skewed. Based on data for 1996-2005, the largest number of recipient countries received aid to the order of 1.8 per cent of their GNI (the mode) per year and 50% of countries received aid at a rate of 3.2 per cent of GNI or less (the median). This corresponds to a distribution of aid per capita with a mode of US$17.9 per year and a median of US$31.5. With this background, the next section turns to the issue of analysing the overall impact of past aid on growth.

4. Does aid support economic growth?

This section addresses a key policy question: does foreign aid raise economic growth rates in developing countries? As discussed in Section 2, economic growth is fundamental to achieving nearly all development objectives. Although the above question may be simple, it is not easy to answer. This stems from methodological challenges as well as weaknesses in the available data. To properly measure the effect of aid, ideally the analyst would like to compare the value of a chosen indicator, growth in our case, in two strictly independent situations – with and without aid. To establish the ‘true’ measure of aid impact, the importance of all other circumstances that have affected growth in a given country over time must be properly accounted for. As we do not simultaneously observe the same country ‘with aid’ and ‘without aid’, the fundamental evaluation challenge is to compare what actually happened with an appropriate counterfactual. In observational studies, there is no way of addressing this problem without making assumptions that are bound to be debatable, both in theory and in practice. For this – and ideological – reasons, the impact of aid remains highly controversial.

The past decades have witnessed a massive outpouring of studies on the effectiveness of foreign aid. Many development experts, subscribing to different paradigms of development

8 thinking, have grappled with this topic; and methodologies have varied. More specifically: (i) the impact of aid has been evaluated at the micro- and macroeconomic levels; (ii) cross- country comparisons as well as single-country case studies have been used; and (iii) broad surveys of a qualitative and interdisciplinary nature, as well as more ‘hard core’ quantitative work, has been pursued. A brief survey of the main findings of this large literature is attempted in the following sub-sections

4.1. Microeconomic evidence

Whilst this note is chiefly concerned with the impact of aid on aggregate indicators, it is useful to begin with the microeconomic evidence (mainly from specific projects). This evidence paints a reasonably positive picture. The most rigorous evaluations in this area are done by the World Bank, and reports from the Independent Evaluation Group (IEG) of the World Bank are encouraging. Average rates of return to aid are generally above 20 per cent, and decent project rates of return have over the years been reported regularly in one survey after another. Overall, a mass of project based evidence has been collected, and few dispute that aid interventions have worked in helping improve outcomes such as better health, improved access to education and promotion of appropriate agricultural technology (i.e., the Green Revolution).

The evaluation of development interventions at the microeconomic level has been bolstered by the recent and rapid expansion of randomized program evaluation techniques, including experimental approaches (for an overview see Langbein, 2006). These seek to avoid potential sources of bias from earlier project evaluations, such as selection bias from choosing only completed (successful) projects, and rely on detailed data which is collected both before and after interventions. Moreover, to address the problem of confounding factors, random selection of beneficiaries is also often used in a similar way to clinical trials of new drugs. Evidence from these studies is also largely positive, indicating that aid-financed interventions can generate substantial benefits for individuals. For example, randomized control trials of providing deworming treatments to school-aged children in Kenya have shown that highly cost effective drugs can reduce school absenteeism by 25 percent and increase school participation by at least 0.14 years of schooling per treated child.

Despite the above, two fundamental challenges remain. First, it is problematic and often inappropriate to generalise the results from a specific evaluation to other contexts. In many cases, it is unclear how a particular effect was generated or what factors may constrain a

9 putative causal mechanism from operating in other situations (for further discussion see Deaton, 2009). Second, neither the results nor the techniques of (randomised) microeconomic evaluation can be directly translated to the aggregate level. The deworming treatments mentioned in the preceding paragraph are a good example. Microeconomic evaluations establish that the deworming programs studied increased the quality (through reduced absenteeism) and quantity (through increased years of schooling) of education. However, many years of continuous and broad scale effort are required to significantly augment the average education level of the working age population. Hence, the growth benefits of the intervention are contingent upon continued success of scaled up programs and even then will only materialize through time.

Rather than attempt to aggregate the impact of numerous microeconomic interventions, the preferred approach to evaluating the macroeconomic impact of aid examines total aid flows and the economic growth rate. Normally, this has been done across countries with the expectation that, other things being equal, countries that receive more assistance would grow more rapidly. The prima facie evidence seems to suggest that countries that have received the most aid have also performed relatively badly in terms of economic growth. This suggests a contradiction between positive microeconomic evidence and the disappointing aggregate performance of countries that have received substantial aid inflows, a paradox often referred to as the micro-macro paradox.

4.2. Three previous generations of macroeconomic studies

A large number of studies have attempted to shed light on this apparent paradox. Previous studies can be classified into three generations, each influenced by dominant theoretical paradigms as well as available empirical tools. The first two generations were inspired by relatively simple models of the growth process such as the Harrod-Domar model and the two- gap Chenery-Strout extension (for exposition see Hansen and Tarp, 2000). The underlying idea behind the Harrod-Domar model is of a stable linear relationship between growth and investment in physical capital. Assuming all aid is invested, it is straightforward to calculate how much aid is required to achieve a target growth rate. The impact of aid is assumed to be positive and helps plug either a savings or a foreign exchange gap. Empirical studies in this tradition consequently focussed on the extent to which aid increases savings and investment in recipient countries. As the detailed survey in Hansen and Tarp (2000) testifies, first generation studies generally concluded that aid does tend to increase total savings, but not by

10 as much as the aid flow. Quite reasonably, this simply suggests a non-negligible proportion of aid is consumed rather than invested.

Retaining the focus on capital accumulation, the second generation of literature moved on to explore the impact of aid on growth via investment. Using data for a cross section of countries, a large number of studies of this kind were produced during the 1980s and early 1990s. Hansen and Tarp (2000) conclude that the findings from these studies consistently indicate a positive link between aid and investment. While a majority of the aid-growth studies of this generation also suggested a positive impact, the result that captured attention was Paul Mosley’s “micro-macro” paradox. This puzzle raised doubts concerning the appropriateness of the underlying growth model and the empirical techniques used. Indeed, it is a tall order to expect both a constant output-capital relationship and that all aid is invested. A second line of critique of the Harrod-Domar and two-gap approach is the argument that growth is less related to physical capital investment than often assumed (Easterly, 1999, 2003). If the productive impact of aid depends more on incentives and relative prices, as well as the policy environment more generally, then it becomes important to consider these broader effects.

The second generation of studies also introduced the problem that poorly performing countries may receive more aid precisely because of their poor growth performance (e.g., Mosley et al. , 1992). There is no necessary logical inconsistency between low growth and high aid inflows. When countries have done well for a while, average incomes will have increased and donors tend to transfer less aid (at least relative to GNI) and may eventually withdraw. At the same time, donor nations have specifically sought to direct aid towards the poorest countries. This implies that, over time, there is a marked tendency for the best performing countries to receive less aid relative to GNI and for the worst performers to receive more. Empirical analyses that do not account for the two-way relationship between aid and growth (that is, the endogeneity of aid) will not reveal aid’s ‘true’ impact. Most second generation studies, however, did not deal with this issue.

From the early 1990s a third generation of more sophisticated econometric studies came to dominate the academic and public discourse about aid. This was motivated by the availability of much better data, allowing analysts to look at changes both across and within countries over time (i.e., panel data became available). Insights from new theories of economic growth, as well as a rapidly increasing numbers of general empirical growth studies, also influenced

11 the research literature. Mindful of the weaknesses of previous studies, the aid-growth relationship came to be perceived as (possibly) non-linear and the endogeneity of aid was taken more seriously. Among the numerous studies of this generation, a leading paper that also came to exert a significant influence on aid policy is the study out of the World Bank by Burnside and Dollar (1997, 2000). The authors argued that although aid has no impact on growth on average, it can work as long as recipients pursue ‘good’ policies. In their own words: “… aid has a positive impact on growth in developing countries with good fiscal, monetary and trade policies ... [but] ... in the presence of poor policies, aid has no positive effect on growth” (2000: 847).

These results, however, were subject to substantial criticism and were shown to be highly fragile. For example, Hansen and Tarp (2001) found that a story of diminishing returns to aid, captured by a squared aid term, best captures the non-linear relationship between aid and growth and is the empirical specification with most support in the data. They concluded that aid has a positive impact on growth but with diminishing returns. Alternatively, Dalgaard et al. (2004), found that variation across countries in the returns to aid seems to be related to their geographical location. Specifically, aid was found to be far less effective in tropical areas over the last 30 years. They also stress, however, that it is hard to believe that aid should be inherently less potent in the tropics. Thus, the real explanation for the aid-tropics link is likely to be elsewhere and the authors call for further research to help disentangle the channels through which aid matters for productivity and efficiency. In an empirical review of many of these contributions, Roodman argues that the results of this generation are extremely sensitive to methodological choices. He goes on to conclude that while some aid is likely to increase investment and growth, aid “is probably not a fundamentally decisive factor for development” (2007: 275). Moreover, due to the multiple kinds of aid, and differences in the efficiency with which it may be put to use, the noise in the data may mask any valuable information regarding the ‘true’ impact of aid.

4.3. Recent contributions

Despite the methodological difficulties, a series of recent studies – a fourth generation – have taken a decidedly pessimistic stance. In particular, in their widely cited paper, Rajan and Subramanian (2008) conclude that aid has had no systematic effect on growth and assert that this conclusion holds across methodologies, time periods and forms of aid. Thus, and as advocated by the authors (2008: 660), an appropriate research agenda may be to unpack the

12 different effects of aid on growth by focussing on intermediate outcomes which are proximate determinants of income growth.

A number of scholars have interpreted this agenda as being one of explaining why resource transfers in the form of aid, at least some of which are invested, don’t lead to an overall increase in the growth rate. A leading explanation for a negative effect of aid that offsets the benefits of resource transfers is that aid undermines or weakens governance by increasing the returns to corruption and/or increasing rent seeking activities. Djankov et al. (2008), for example, argue that aid has a statistically significant negative effect on changes in political institutions (specifically, democracy). Aid is likened to natural resource windfalls, which in principle could be used to boost growth, but in practice have often resulted in poor governance and a poor growth record over time. According to Djankov et al. (2008), the negative effect of aid on political institutions is larger in magnitude than that caused by natural resource windfalls. Similarly, in an earlier study, Rajan and Subramanian (2007) find that manufacturing sector growth in developing countries is undermined by a detrimental effect of aid inflows on governance.

Other scholars argue that, given that the association between income growth and investment in both human capital (skills, education, health) and physical capital appears to be fragile within countries over time (Kenny, 2005), it is unsurprising that a robust relationship between aid and growth has been elusive. This accords with the view that growth is an extremely complex, idiosyncratic and open phenomenon that cannot easily be captured by simple empirical models (see Temple, 2000). Indeed, in the wider development economics literature, there has been a shift away from grand theories of how developing countries grow, towards a more nuanced emphasis on binding constraints (Rodrik, 2007) and varieties of coordination challenges. All-purpose explanations of growth that apply equally across countries over time are viewed with scepticism. Collier (2007), for instance, identifies numerous growth traps to which low income countries may be subject and each of which demands a different policy response. Accordingly, he argues that foreign aid is only likely to directly boost growth in a relatively small number of cases. In the remaining cases, it will often only sustain countries on a form of minimal life support.

In light of this increasing humility concerning what we know about how growth is initiated and then sustained, a set of recent papers focuses on the effects of aid on non-growth outcomes such as infant mortality, health, and education. Mishra and Newhouse (2007), for

13 example, find a small but statistically significant effect of health aid on reductions in infant mortality. Masud and Yontcheva (2005) also find that some types of aid help reduce infant mortality.

Recent theoretical exercises give further weight to the need for modest expectations regarding the magnitude of the possible impact of foreign aid on growth. In particular, Rajan and Subramanian (2008) have contributed to this literature. Based on a standard neoclassical production function framework, and assuming that aid only increases physical capital investment, they estimate that the overall increase in the growth rate accruing from aid inflows of 10% of GDP may only lie in the range of 1% to 2.5% depending on the share of aid that is invested and productivity impacts. Thus, at the levels of aid seen by most countries, these effects may be very difficult to distinguish from business cycle fluctuations and external shocks, especially when endogeneity of aid and the quality of data on GDP and aid flows are considered. In fact, while Rajan and Subramanian emphasize that their estimates rarely have an effect on growth that is statistically different from zero, they are also unable to conclude statistically that the impact of aid on growth falls outside of the range that they suggested. This is particularly true when longer time frames are considered. In these cases, the estimated parameters are much closer to the suggested ranges than they are to zero.

Furthermore, the range proposed by Rajan and Subramanian may be overoptimistic. Using a more sophisticated theoretical approach, Dalgaard and Erickson (2009) find that expectations with respect to the impact of aid on growth should be even more modest. They consider aid inflows to sub-Saharan Africa over the past 30 years, which averaged around 5% of recipient GDP, and calculate an expected annual increase in the growth rate from past aid flows of around 0.1%. This assumes that all aid is invested and there are no productivity impacts. An effect of this magnitude would be impossible to isolate from other factors affecting growth with existing data and methods. Whatever the expected reasonable range, a fundamental point is that, once one uses more realistic theoretical models compared to those used in the first and second generation literatures (see Easterly, 1999), the expected potency of aid in stimulating growth declines. Given the ranges predicted by modern growth models, there may be no aid effectiveness puzzle after all.

Finally, research attention also has been given to long-run determinants of growth that have a cumulative but often not immediate impact on the rate of income growth. Changes in human capital, such as education and health, move only slowly at the aggregate level and exert a

14 positive influence on economic growth with a substantial lag. This comes from simple demographics whereby improvements in schooling indicators, for example at the primary level, can take many years to translate into noticeable increases in average education levels among working age adults. Ashraf et al. (2008) demonstrate that the immediate economic impact of gains in life expectancy from disease eradication may be a reduction in per capita incomes due to increased child survival and the consequent increase in the ratio of the non- working age to the working age population. In short, as a result of these interventions, incremental growth in population can easily exceed the incremental growth in income for two, three or even more decades following interventions.

The impressive performance of many developing countries, including countries in sub- Saharan Africa, in improving social indicators such as expanding access to education, reducing infant mortality, and expanding access to health services over the past three decades may only now be translating into gains in per capita income. This reinforces the need for modest expectations regarding the effect of aid on observed income growth. Where aid has been used primarily to boost social spending, as advocated under the MDGs, it may be inappropriate to investigate the aid-growth relationship under anything other than a time horizon of 30-40 years.

In response to these concerns, there have been calls for more detailed country case studies, which permit a more nuanced assessment. Arndt et al. (2007) provide a case study of Mozambique in which they attempt to evaluate the effects of aid across different proximate drivers of growth. Starting with long-run growth accounting estimates, they find that aid has played a crucial role in rebuilding infrastructure and expanding access to health and education. Aid has supported rapid reconstruction and seems to have crowded-in private and foreign investment, but has also brought substantial governance and economic management challenges. This provides supporting evidence for Collier’s (2007) argument that aid can be particularly beneficial in post-war environments but also brings new challenges.

Finally, in their recent up-to-date cross-country study Arndt, Jones and Tarp (2009) note that while the micro-macro paradox seems to have been revived, a balanced view of the evidence is that aid remains an important tool for enhancing the development prospects of the poor. Arndt et al. apply micro-econometric techniques from the program evaluation literature and lessons from the growth literature in a novel way, and their findings show that aid has a positive and statistically significant causal effect on growth over the long run, with point

15 estimates at levels suggested by growth theory. Arguably, their approach represents the most carefully developed empirical strategy employed in the aid-growth literature to date, and the average treatment effect of aid on growth is positive in both the 1970-2000 and 1960-2000 periods. In fact, the statistically significant point estimates suggest that an inflow on the order of ten percent of GDP spurs per capita growth by more than two percent per annum in the long run. These estimates are consistent with the view that foreign aid stimulates aggregate investment and also contributes to productivity growth, despite some fraction of aid being allocated to consumption. The 95% confidence interval around these estimates lies in the strictly positive domain and contains the prior, suggested in RS08 by Rajan and Subramanian (2008) that the long-run elasticity of growth to foreign aid should be 0.1. In the shorter term, however, the analysis by Arndt et al. indicates that the impact of aid is difficult to discern. Nevertheless, combining the longer run macro evidence with the evidence at the micro- and meso-levels, a consistent case for aid effectiveness emerges.

5. Some stylized lessons

Despite what may be seen as the rather paltry fruits of hundreds of academic studies, a number of lessons have been learnt. These can be summarised as follows: a. We now have a better understanding of the difficulties involved in investigating whether aid stimulates growth on aggregate. In particular, we have a solid grasp of why empirical results are often biased towards zero (even if there is a positive true long-run relationship). The principal factors behind this are noisy and weak data; the high complexity and openness of the growth process; the endogeneity of aid; and the demographic implications of gains in social indicators that may be driven by aid. Furthermore, methods that enable the analyst to deal with all of these factors simultaneously, and thereby tease-out the ‘true’ impact of growth, are only beginning to emerge. b. Putting these methodological difficulties aside, it is essential that one maintains appropriately modest expectations about the magnitude of aggregate returns to aid. As discussed above, theoretical exercises built on realistic assumptions tend to indicate small growth gains from the average volumes of aid that have been donated in the past (and which are likely to continue). Such expectations are significantly smaller than those generated from the early generations of research on aid and growth. However, this

16

message is not widely disseminated and needs to be absorbed by the wider policy community and general public. c. It is important to recognise the multiple channels through which aid may have an impact on growth, as well as corresponding differences in the horizon over which aggregate effects may become apparent. Improvements in health, education and institutions can take a very long time (up to a generation) to cumulate into potential growth effects. This indicates that non-growth aggregate outcomes associated with aid, such as changes in education or health, can be usefully studied in their own right. At the same time, this also means that countries with very different initial conditions and/or aid profiles may see very different responses to aid over time. Few studies, if any, have rigorously and explicitly taken this issue into account. d. Finally, an obvious but often forgotten point is that history matters. Aid has not been provided on a random basis, nor has there been any kind of stability in donor aid allocation strategies or views about what constitutes the ‘right’ set of policies to induce growth. Consequently, one must exercise substantial caution in using past trends to inform about future possibilities. There is no doubt that during the Cold War, aid was often used as a geo-political tool of advanced countries. Today, the emphasis is on poverty, social welfare and conflict (or terrorism) prevention. Of course these shifts in donor behaviour and global conditions further frustrate the methodological challenge, but they cannot be ignored. The point that political motives largely guide the size and the form of aid is of crucial importance. Clearly if the motivation of donors is not growth and development then it is not that surprising that the link between aid and growth is tenuous. A related point is that the lack of coordination among donors and sometimes conflicting objectives further reduce the impact of aid on growth. Still another issue – relevant in a number of circumstances including recent Chinese aid via-à-vis SSA is aid provided as a carrot to have access to natural resources.

So, where does this leave us? The tone of the very recent literature is pessimistic, this literature comes nowhere near to confining the estimated causal impact of aid on growth to be less than the range reasonable expectations would suggest. Moreover, the pendulum in the literature tends to swing slowly with more critical assessments tending to follow more positive assessments and so forth. At the same time, methodological improvements are being realized by works on both sides of the debate. As shown in the study by Arndt, Jones and

17

Tarp (2009), “recent literature” is now introducing new methodological improvements that arrive at a more positive tone.

Overall, looking broadly across the full range of existing literature from microeconomic evidence to the cross country macroeconomic analyses relating aid to growth, the balance of evidence indicates positive effects, particularly when longer time frames are considered. In this, Collier argues that over the past 30 years aid has probably increased the annual growth rate of the poorest countries by around one percentage point and adds that: “Without aid, cumulatively the countries of the bottom billion would have become much poorer than they are today” (2007: 100). These conclusions do not deny that aid can be inefficient or that negative side-effects may arise from specific projects or relationships. Nor do they imply that we should stop learning about how aid works or how to do aid better. Here, on the contrary, much work remains to be done.

6. Conclusion

Arndt, Jones and Tarp (2009) begin their recent paper by noting that the extent to which foreign aid can be a decisive factor in the economic development of low income countries remains controversial. In 1987, Paul Mosley suggested that while aid seems to be effective at the microeconomic level, any positive aggregate impact of aid is much harder to identify (Mosley, 1987). He labeled this the micro-macro paradox, and it challenged the conclusions of the seminal works by Papanek (1972, 1973). After more than twenty years, Rajan and Subramanian (2008) concluded “it is difficult to discern any systematic effect of aid on growth”. At the same time, microeconomic evaluations, including rigorous contributions to the program evaluation literature by development economists, have demonstrated the potential for well-designed interventions to achieve significant positive results.

Thus, after two decades of intense analytical work using new theory, new data and new empirical methodologies, it might appear that the micro-macro paradox has been revived. Other similarities with the late 1980s and early 1990s exist, not least with respect to policy. In 1994, the Economist magazine concluded from the results of Boone (1994) that “Aid [goes] Down the Rathole”. Today, in the midst of a serious global economic crisis, where aid is arguably more needed than ever, the attention of both the aid community and decision- makers is on “Dead Aid” (Moyo, 2009), which argues for a complete cessation of aid flows to Africa.

18

So, has the aid and growth literature come full circle? The response of Arndt, Jones and Tarp (2009) is “no”. While the pendulum has swung to skepticism concerning the ability of aid to contribute to economic growth in the most recent literature, a series of important points of agreement have emerged. First, methodological advances have improved the profession’s capacity to identify causal effects in economic phenomena. These advances in methods are beginning to be applied at the more aggregate level; and in this regard, the supply side instrumentation approach of Rajan and Subramanian counts as a significant advance (with room for improvement). Second, these methodological advances highlight the serious challenges that must be surmounted in order to derive robust causal conclusions from observational data. In many important areas of inquiry, longstanding debates with respect to causal impacts persist despite improved methods and improved data availability. Third, the formation of reasonable expectations about the likely returns to foreign assistance has been greatly facilitated by the application of modern growth theory. Finally, there is an increasing recognition that many of the key interventions pursued by foreign aid will only result in positive growth outcomes over long time horizons.

Arndt, Jones and Tarp (2009) also show that the average treatment effect of aid on growth is positive in both the 1970-2000 and 1960-2000 periods. The long run elasticity of growth with respect to the share of aid in recipient GDP appears to around 0.20. This is consistent with the view that foreign aid stimulates aggregate investment and also may contribute to productivity growth, despite some fraction of aid being allocated to consumption. In sum, when combined with the evidence at the micro- and meso-levels, a consistent case for aid effectiveness emerges. There is no paradox. Overall, the bleak pessimism of much of the recent aid-growth literature is unjustified and the associated policy implications drawn from this literature are often inappropriate and unhelpful. Aid has been and remains an important tool for enhancing the development prospects of poor nations.

Finally, nearly all participants in the aid-growth debate recognize the potential for aid to do better, particularly in fostering productivity growth. The evidence indicates that sustaining foreign assistance programs at reasonable levels can be expected to enhance the living standards of the world’s poorest people. Abolishing foreign aid, or drastically cutting it back, would be a mistake and is not warranted by any reasonable interpretation of the evidence. The challenge is to improve foreign assistance effectiveness so that living standards in poor countries are substantially advanced over the next three decades. And that it can be done is

19 evident from historical examples such as Taiwan, Korea and Mozambique where aid has made major contributions to growth and development.

20

References

Alesina, A. and D. Dollar (2000), Who Gives Foreign Aid to Whom and Why?, Journal of Economic Growth , 5: 33–63 Arndt, C., Jones, S. and Tarp, F. (2009), ‘Aid and Growth: Have We Come Full Circle?’ WIDER Discussion Paper No. 2009/05 Arndt, C., Jones, S. and Tarp, F. (2007), ‘Aid and Development: the Mozambican Case’, in Lahiri, S. (ed.), Theory and Practice of Foreign Aid , Frontiers of Economics and Globalization, Volume 1, Amsterdam: Elsevier Ashraf, Q. and O. Galor (2008), Dynamics and Stagnation in the Malthusain Epoch: Theory and Evidence, Brown University Working Paper , 2008-14, Department of Economics Boone, P. 1994. “The Impact of Foreign Aid on Savings and Growth.” London School of Economics CEP Working Paper 1994 (677) Burnside, C. and D. Dollar (2000), Aid, Policies, and Growth, American Economic Review 90: 847-68 Collier, P. (2007), The Bottom Billion , Oxford: Oxford University Press Dalgaard, C. and L. Erickson (2009), Reasonable Expectations and the First Millennium Development Goal: How Much Can Aid Achieve, World Development , forthcoming Dalgaard, C.-J., H. Hansen and F. Tarp (2004), On the Empirics of Foreign Aid and Growth, The Economic Journal , 114: F191-216 Deaton, A. (2009), Instruments of Development: Randomization in the Tropics, and the Search for the Elusive Keys to Economic Development, NBER Working Paper , 14690 Djankov, S., J. Montalvo and M. Reynal-Querol (2008), The curse of aid, Journal of Economic Growth , 13(3): 169-194 Easterly, W. (1999), The Ghost of Financing Gap: Testing the Growth Model Used in the International Financial Institutions, Journal of Development Economics, 60: 423-38. Easterly, W. (2001), The Elusive Quest for Growth , MIT Press: Cambridge MA. Easterly, W. (2003). Can Foreign Aid Buy Growth?, Journal of Economic Perspectives 17: 23-48 Easterly, W. (2006), The White Man’s Burden: Why the West’s Efforts to Aid the Rest Have Done so Much Ill and so Little Good , Penguin Press: New York Easterly, W. (2009), How the Millennium Development Goals are Unfair to Africa, World Development, 37(1): 26–35 Hansen, H., and F. Tarp (2000), Aid Effectiveness Disputed, Journal of International Development , 12: 375-98

21

Hansen, H., and F. Tarp (2001), Aid and Growth Regressions, Journal of Development Economics , 64: 547–70 Kenny, C. (2005), There’s (Much) More to Life than Income: The Weak Link Between Income, Health, and Education, Unpublished report, World Bank Langbein L. (2006), Public Program Evaluation: A Statistical Guide , M.E. Sharpe

McGillivray, M. (2003), Modelling Foreign Aid Allocation: Issues, Approaches and Results, Journal of Economic Development , 28(1): 171-188 Moyo, D. 2009. Dead Aid: Why Aid Is Not Working and How There Is a Better Way for Africa . London: Allen Lane Mishra, P. and D. Newhouse (2007), Health Aid and Infant Mortality, IMF Working Papers, 07/100, International Monetary Fund Mosley, P. 1987. Overseas Aid: Its Defence and Reform. Brighton: Wheatshead Books Papanek, G. F. 1972. “The Effect of Aid and Other Resource Transfers on Savings and Growth in Less Developed Countries.” Economic Journal 82 (327): 935-50 Papanek, G. F. 1973. “Aid, Foreign Private Investment, Savings, and Growth in Less Developed Countries.” Journal of Political Economy 81 (1): 120-30 Rajan, R.G., and A. Subramanian (2007), Does Aid Affect Governance?, American Economic Review , 97(2): 322-327 Rajan, R.G., and A. Subramanian (2008), Aid and growth: what does the cross-country evidence really show?, The Review of Economics and Statistics , 90(4): 643–665 Riddell, R. (2007), Does Foreign Aid Really Work? , Oxford University Press: Oxford Rodrik, D. (2007), One Economics, Many Recipes: Globalization, Institutions, and Economic Growth , Princeton University Press Roodman, D. (2007), The Anarchy of Numbers: Aid, Development, and Cross-Country Empirics, The World Bank Economic Review , 21(2): 255–277 Sachs, J. (2005), The End of Poverty: Economic Possibilities for Our Time , Penguin Press: New York Tarp, F. (2006), Aid and development, Swedish Economic Policy Review , 13(2): 9–61

Temple, J. (2000), Growth Regressions and What the Textbooks Don't Tell You, Bulletin of Economic Research , 52(3): 181-205 Yontcheva, B. and N. Masud (2005), Does Foreign Aid Reduce Poverty? Empirical Evidence from Nongovernmental and Bilateral Aid, IMF Working Papers 05/100, International Monetary Fund

22